The Journal of the American Chemical Society with which has been incorporated
The American Chemical Journal (Founded by Ira Remsen)
VOL. 46
MAY, 1924
[CONTRIBUTION FROM
THE
No. 5
LABORATORY OF PHYSICAL CHEMISTRY OF THE UNIVERSITY OF WISCONSIN]
ADSORPTION AND SURFACE TENSION AT LIQUID-LIQUID INTERFACE BY J. HOWARD MATHEWS AND ALFRED J. STAMM R ~ C B I V EOCTOBER D 8, 1923
The determination of the extent of surface is the chief difficulty confronted in adsorption problems. This has, to some extent, been solved for solid adsorbents by Fritz Paneth and Walter Vonverkl but their method is limited to salts of metals which have radioactive isotopes. The extent of surface can more readily be determined for adsorption a t liquidgas and liquid-liquid interfaces. I,angmuir2 has taken advantage of this fact in his investigations on spreading of oils from which he developed, with the aid of material from other sources, his important theory of adsorption. Harkins and his co-worker$ have done considerable work on the structure of the surface of liquids from the standpoint of surface energy obtained from surface tension and other similar data. Their results confirm the theory of Langmuir, as do the investigations of Adam4 on “The Structure of Thin Films of Palmitic Acid on Water” and those of Iredale5 on “Adsorption from the Gas Phase a t a Liquid-Gas Interface.” The work described in this paper deals primarily with the application of the Gibbs adsorption equation, which gives the increase in moles per sq. cm. in the interface of a system. Paneth and Vorwerk, Z. physik. Chem., 101, 446 (1922). (a) 37, 1154 (1915); (b) 38, 2221 (1916); (c) 39, 1848 ( 1917). Harkins and others, ibid., 39, 354, 541 (1917); 41, 970 (1919); 42, 700 (1920); 43, 51 (1921); 44,2665 (1922). Adam, P70C. Roy. SOC.,99A, 336 (1921). Iredale, Phil.Mag.,45, 1088 (1923).
* Langmuir, THISJOURNAL,
1072
J. HOWARD MATHEWS AND ALFRED J. STAMM
Vol. 46
This equation has, since Gibbs’ time, been derived in several different ways, and has been developed by Milner6 without assuming that the perfect gas law holds in the surface layer. It is necessary only that it should hold for the osmotic pressures within the liquids. The experimental verification of this relationship is very difficult and has been undertaken only by a few experimenters, Donrian and Baker’ successfully confirmed the relationship for adsorption a t a liquid-gas interface. W. C. McC. Lewiss attempted similar determinations for adsorption a t the liquid-liquid interface of a hydrocarbon oil and water, with less success. The observed values were in almost all cases a great deal larger than the calculated. There are, however, a great many sources for discrepancies and difficulties connected with the determination which W. C. McC. Lewis recognized, and he himself does not consider the results as disproving the relationship. Harkins and his co-workersg have developed the technique of surfacetension determinations by the drop-weight method and have applied it with success to the determination of interfacial tension. Morgan and his co-workerslO have also done a great deal of work on the drop-weight method. Harkins’ work is undoubtedly the most complete. He has determined the correction factor for the residual drop by careful experiments with the capillary-rise method. Further, he has shown that these corrections are equally applicable to interfacial tension, by measurements of capillary rise a t liquid-liquid interface. The correction curve for residual drop was made from the tables of data of Harkins giving the relationship
of (r/V”’) tof
ma (9 .~
where r is the radius of tip and V the volume of the drop.
The equation for interfacial surface tension in dynes per cm. is
f
=
where pw and P O are the densities of water and the other
vd(pw-po)
2 n- r N
CT
(6) .
.
liquid, respectively. The apparatus devised for this work is shown in Fig. 1. The pipet consists of a bulb of 25.74 cc. * 0.005 capacity, joined by (0.5 cm. i. d.) tubing to the capillary tip (of 2 cm. i. d.). The inside bore was ground out to a distance of approximately 0.5 cm. from the end to a diameter of 0.351 cm. The inside diameter of the tip was determined by turning a brass rod to fit snugly into the bore of the tip. The diameter of the brass rod was measured with micrometer calipers, and the average of a number of readings taken. These did not differ by more than 0.002 cm. The Milner, Phil. Mag., 13, 96 (1907). Donnan and Baker, Proc. Ray. Soc., 85A, 557 (1911). 8 Lewis, Phil. Mug., 15,499 (1908); 17,466 (1909); Z . physik. Chem., 73, 129 (1910). 9 Harkins and others, THIS JOURNAL, 38, 228 (1916); 41,499 (1919). 10 Morgan and others, ibid., (a) 37, 1461 (1915); (b) 38, 844 (1916); (c) 39, 2151, 2275 (1917). 6 7
.
May, 1924
ADSORPTION
1073
BT LIQUID-LIQUID INTERFACE
upper part of this pipet was provided with two tubes closed by stopcocks, one for filling and the other with a fine orifice to control the displacement of liquid by air. An automatic counter was devised to count the drops. In its final form it worked very well and proved dependable, as shown when checked by actual counting and by the reproducibility of results. The device consisted essentially of a platinum vane of thin foil balanced on a platinum pinion so that when a drop of the liquid on being released from the tip impinged on it, the pointer on the other end dipped into the cup of mercury thus making contact. The terminals were connected to the secondary of a relay, with a switch and three storage batteries in series. The primary of the relay was connected in series with the counter and the batteries. The contact points of the relay were interchanged so that instead of one side breaking as the other made contact, they both made contact and broke together. The counter consisted of an electromagnet which operated a Veeder ratchet counter. When a drop impinged on the platinum vane and lifted it slightly, the circuit was completed and the relay operated the electromagnet. The armature of the electromagnet operated the ratchet lever of the counter, causing the next number to be brought into position. The benzene was purified according to the standard method, to remove thiophene. A final fraction boiling a t 79.5-79.7' (750.7 mm.) was obtained. The dimethylaniline used was purified by distillation, collection of the fraction between 190" and 191", freezing it twice to remove methylaniline, and again distilling it. The fraction used distilled a t 190.5-190.8' (740.4 mm.). The heptane was secured through the kindness of Professor Kremers of the Pharmacy Department of the University of WisFig. 1. consin. The interesting study of its production and properties has been covered by Professor Kremers in a paper on "Chemistry of Heptane and its Sohtion."*l The sample used distilled between 98.25' and 98.40' a t 751 mm.
The densities of the materials were determined with a Sprengel pycnometer a t 25' and were referred to water a t the same temperature. The densities of several mixtures were determined to see whether the substances followed the mixture law, with the following results. TABLE I DENSITIESAT 25' REPERRED TO WATERAT 25'
Benzene (c. P. but not purified as above). . . . . . . . . . . . . . . . . . . 0,8753 Dimethylaniline.. . . . . . . . .9565 .6824 Heptane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benzene-dimethylaniline, 50-50 (by vol.) ob?. . . . . . . . . . . . . . .9162 Heptane-dimethylaniline,
50-50 (by vol.) obs. .
0.04%
........
.8210
.IS% I1
J . Ant. Pharm. Assoc., 9, No. 9, 857 (1920); 10, No. 11 (1921).
1074
Vol. 46
J. HOWARD MATHEWS AND ALFRED J. STAMM
According to Brown1$ and Linebarger,13 who made a number of determinations of densities of mixed liquids, the maximum deviation is near the mixture of equal proportions. As the above deviations are within the experimental error of the rest of the work, the calculated densities for mixtures have been used in this paper. All of the surface-tension determinations were made in a thermostat held a t 2 5 O * 0.02'. The pipet was placed in a glass cylinder 7.5 cm. in diameter and held in place with a specially cut cork. The cylinder was filled to a height of 11.5 cm. with distilled water, except when otherwise noted. The pipet was filled with the original liquid to a point a little above the upper line, placed in position in the thermostat and allowed to stand for 20 minutes to half an hour, in order to come to the temperature of the thermostat. In starting a run, the rate of dropping was brought to the desired point by the time the second drop was released. This could be held quite constant with about two adjustments of the stopcock during the run to compensate for loss in head. The tables of data show that different runs could be checked to one-half a drop, representing an accuracy of from 1% to less than 0.5%. Test runs were made by first TABLE I1 NO.
1 2 3 4 5 6 7 8
1 2 3 4 5 6
7 8 9 10 11 12 13 14 l2
13
DIMETHYLAXILINE-HEPTANE AGAINST WATER AT yo Dimethylaniline by vol. d:: Drop. no.
25
0,6825 205.75 0 .OO I. .19 ,6858 222 . 0 ,7021 248.5 7.14 14.30 ,7216 255 . 0 ,7509 246 .O 25 .00 ,8194 189.25 50 .OO ,8878 125..5 75 .OO ,9563 50.80 100 .OO DIMETHYLANILINE-BENZENE AGAINST WATER 0 .OO 0.8758 112.05 3 .00 .8780 112.75 6.22 ,8808 112 . 0 7.14 ,8816 111.5 ,8847 111.25 11.10 ,8870 110.0 13 .90 110 . 0 16.70 ,8892 106 .0 22.70 .8941 ,9026 99.5 33.33 89.37 50.00 ,9160 83.75 57.20 ,9218 .9333 73.37 71.50 ,9429 64.37 83.30 50.80 100.00 ,9563
Brown, J . Chem. SOL,39, 202 (1881). Linebarger, Am. Chem. J., 18,429 (1896).
Surface tension (dynes per cm.)
50.39 46.48 39,65 36.22 33.49 30.98 28.23 25.57 34.71 33.93 33.33 33.25 32.45 32.12 31.50 31.15 30.40 28.98 28.68 27.65 26.78 25.57
May, 1924
1075
ADSORPTION AT LIQUID-LIQUID INTERFACE
saturating the water with the organic solvent, and without saturating it. The results checked in all cases within the experimental error; hence saturation of the water was not resorted to. A series of runs was made with benzene to determine the maximum rate of drop formation which Dimethylaniline-heptane against water.
50
8 45 40 m
E35
n
30 25 0
10
20
30
40 50 60 Vol. per cent. Fig. 2.
70
80
90
100
would give results comparable with static conditions. This was found to be six drops per minute. Varying the head in the cylinder by 2 cm. of water was found not to affect the results. The surface tensions of mixtures of two pure organic liquids against water, which themselves have considerably different surface tensions Dimethylaniline-benzene against water,
35
33
31
& $29
dx
27 25
0
10
20
30
40 50 60 Vol. per cent. Fig. 3.
70
80
90
100
against water, were determined. The two systems, dimethylanilinebenzene against water, and dimethylaniline-heptane against water, were studied. The densities of these pairs follow the mixture law, showing that a solution of one in the other is not accompanied by association or
1076
Vol. 46
J. HOWARD MATHEWS AND ALFRED J. STAMM
dissociation. The pairs are also miscible in all proportions. The data for these mixtures are given in Table 11and the curves in Figs. 2 and 3. The drop numbers and surface tensions are averages of several determinations. The rate of dropping in each case was less than six drops per minute. The curves fall below the line joining the tensions of the pure constituents. The constituent of lower tension when added to the higher causes a considerable fall in surface tension, due evidently to the constituent
.4
.8
1 .2 Log c. Fig. 4.
0
.4
of lower tension concentrating in the interface. On the other hand, adding the constituent of higher tension to the lower has less effect, as it goes mostly into the interior of the drop. The data as plotted in Fig. 4, surface tension against the log of concentration in g. per cc., make it possible to determine (I') moles per sq. cm. adsorbed on the interface. Where the slope of the curve becomes constant, da/d log C, the effective amount adsorbed becomes constant due to complete covering of the surface. This was observed by Langmuirl4 and by King,15 and used as a basis of calculating molecular cross sections. King found that butyric acid adsorbed a t water-air and water-benzene interl4 l6
Ref. 2c, p. 1898. King, Kansas Exp. Sta. Teck. BuIl., 9 (1922).
May, 1924
ADSORPTION AT LIQUID-LIQUID INTERFACE
1077
faces had the same molecular cross section, but that a greater concentration of butyric acid was required to give a complete interface layer in the former case. This can be explained on the basis of polarity. Butyric acid, which contains both a polar and a non-polar group, when dissolved in a polar liquid tends to go to the interface and become oriented so that the polar group remains in the polar solvent and the non-polar group, which is repelled, extends upwards into the indifferent air medium. When the interface, however, is water-benzene, in addition to the above forces there is the attractive force of the non-polar benzene for the non-polar group of the butyric acid. This increases the tendency of the butyric acid to go to the interface and lowers the concentration necessary t o form a completely adsorbed layer. The fact that alcohol adsorbed a t waterair interface forms a complete adsorption layer a t a lower concentration than the corresponding fatty acid, can be similarly explained. Each contains a polar and a non-polar group. The acid contains the group COOH, which is more polar than OH, as is shown by the greater reactivity of organic acids than alcohols. It thus has a greater tendency to remain in the polar solvent which opposes the tendency to become oriented a t the interface. According to the calculations of Langmuir on the data of Morgan and Egloffle triethylamine forms a complete surface layer a t water-air interface a t very low concentrations, whereas with phenol a considerably greater concentration is required. As the two hydrogens of the amine have been replaced by non-polar ethyl groups the compound is quite non-polar and tends to leave the polar solvent readily, whereas the phenol which has the polar hydroxyl group tends to remain in the polar solvent to a greater extent. Applying this reasoning to the above data, the dimethylaniline, which is but slightly polar, tends to remain in the non-polar solvent, and hence a considerably higher concentration would be required to give a complete interfacial layer. This agrees entirely with the experimental results. The molecular cross sections for dimethylaniline a t heptane-water and benzene-water interfaces calculated on this basis are 0.695 X 1 0 - 1 4 and 0.758 X an2,respectively. By making two assumptions, the calculations can be made on an entirely different basis in which the thickness of adsorption is obtained directly and the area of cross section determined from this with the aid of Avogadro’s number and the density. The first of these assumptions is that if there were no concentration in this interface, the surface tension of the mixtures would fall on the straight line joining the pure constituents against water, when plotted on a volume-percentage basis. Experiments have ~ the additivity law for surface tension on a been made by W ~ r l e yto~test Ref. 10 b, p. 848. 17Worley,J . Clzem. Soc., 105, 276 (1914).
1078
Vol. 46
J. HOWARD MATHEWS AND ALFRED J. STAMM
volume-percentage basis. For constituents of like polarity against air, which have no orienting effect, the relationship was fairly well followed, as would be expected when there was no concentrating in the interface. The assumption, hence, seems perfectly reasonable for liquids whose solutions show no such abnormalities as association or dissociation, since such other properties as density and refractive index are straight line functions. The other assumption is that it is justifiable to use concentration units down to thicknesses of molecular dimensions, From the first assumption, the percentage composition in the interface corresponding to the percentage composition in the drop can be determined from Figs. 2 and 3 by drawing a horizontal line from the curve to the straight line joining the tensions of the pure constituents against water. By converting the interfacial concentration from percentage of dimethylaniline to moles of dimethylaniline per cc., the excess of dimethylaniline per cc. in the interface over that in the drop can bedetermined by difference, the change in concentration in the drop due to change in concentration in the interface being negligible. The thickness of the adsorbed layer can be calculated by dividing the value of (I?), moles per cmS2,by the TABLE I11
THICKNESS OF DIMETHYLANILINE MOLECULES CALCULATED FROM SURFACE-TENSION DATA Heptane-Water Interface IXmethvlnniiif;e
